The small chromatin-binding protein p8 coordinates the association of anti-proliferative and pro-myogenic proteins at the myogenin promoter

Myogenic differentiation is an ordered process during which determined myoblasts withdraw from the cell cycle and, under instructions from a set of myogenic transcription factors, the expression of muscle-specific genes within these cells is sequentially activated. Differentiation proceeds by fusion of mononucleated precursors into multinucleated myotubes, which assemble the specialized contractile cytoskeleton of skeletal muscle. Whereas cell-cycle arrest is essential for differentiation, muscle cells can also withdraw into a reversible quiescent state in which tissue-specific gene expression is suppressed. These alternate `out-of-cycle' states (permanent vs temporary) might result from coordination vs uncoupling of the programs of cell-cycle exit and differentiation. The mechanisms that regulate this divergence are poorly understood, and are relevant to an understanding of how a balance of precursor and differentiated cells arises during muscle formation and regeneration.

Much of our understanding about the molecular nature of crosstalk between the cell cycle and myogenesis comes from analysis of the C2C12 myogenic cell line (Olson, 1992; Halevy et al., 1995; Andres and Walsh, 1996). Inhibitory interactions between cell-cycle activators and lineage-specific transcription factors underlie the mutually exclusive programs of proliferation and differentiation (reviewed by Wei and Paterson, 2001; Lassar et al., 1994). Of the muscle regulatory factors [MRFs; a family of basic helix-loop-helix (bHLH) transcription factors that orchestrate myogenesis], the determination factors MyoD and Myf5 are expressed in proliferating, undifferentiated myoblasts (reviewed by Kitzmann and Fernandez, 2001), whereas myogenin is induced during early differentiation (Wright et al., 1989; Andres and Walsh, 1996), activated in culture by serum withdrawal. Cell-cycle inhibitors reinforce the action of differentiation-promoting MRFs. For example, induction of the cyclin-dependent kinase inhibitor (CDKI) p21 (a transcriptional target of MyoD) in myogenin-expressing cells (Halevy et al., 1995; Andres and Walsh, 1996) is followed by a cascade of muscle-specific gene activation and fusion into multinucleated post-mitotic myotubes (reviewed by Buckingham, 2001). By contrast, exit of undifferentiated myoblasts into G0 is accompanied by loss of MyoD expression, lack of myogenin and p21 induction, and, consequently, absence of muscle-specific gene induction (Milasincic et al., 1996; Kitzmann et al., 1998; Sachidanandan et al., 2002; Dhawan and Helfman, 2004). G0 is reversible, and the G0-G1 transition in myoblasts is accompanied by reactivation of MyoD expression (Gopinath et al., 2007) (reviewed by Dhawan and Rando, 2005). Thus, cell-cycle re-entry correlates with the re-emergence of the capacity to differentiate, but differentiation itself requires additional signals.

Competence for differentiation is confined to the G1 phase of the myoblast cell cycle (Clegg et al., 1987). This phase, prior to the retinoblastoma (Rb)-controlled restriction point, is the period when cells are responsive to environmental cues (Planas-Silva and Weinberg, 1997). Thus, cells must be in cycle in order to respond to extracellular differentiation signals; arrested cells cannot. The molecular logic for this restriction is that MyoD expression is suppressed in G0 (Milasincic et al., 1996; Kitzmann et al., 1998). MyoD expression is regulated by multiple factors, including Pax3 and Pax7 (reviewed by Buckingham 2001), serum response factor (SRF) (Carnac et al., 1998; L'Honore et al., 2003; Gopinath et al., 2007), and the transcriptional co-activator p300 (also known as CBP), a histone acetyl transferase (HAT) (reviewed by McKinsey et al., 2001; McKinsey et al., 2002). Post-translational modifications of MyoD also modulate its ability to promote differentiation. In particular, p300-HAT-dependent acetylation of MyoD is crucial (Polesskaya et al., 2001; Puri et al., 1997b; Dilworth et al., 2004), whereas deacetylation of MyoD by HDAC1 silences MyoD transcriptional function (Mal et al., 2001).

Although occupancy of its target promoters is the rate-limiting step for transcriptional activation by MyoD, remodeling of chromatin at these sites is also essential for efficient transcription (Gerber et al., 1997; de la Serna et al., 2005). MyoD-HDAC (histone deacetylase) complexes repressing muscle-specific promoters in growing myoblasts are replaced by MyoD-HAT complexes in differentiation conditions, leading to increased accessibility and robust transcriptional activation (Mal and Harter, 2003; Yuan et al., 1996; Puri et al., 1997b; McKinsey et al., 2001). p300-HAT activity is therefore required not only for acetylation of MyoD protein but also for histone acetylation at the target promoters of MyoD.

In this study, we analyzed the expression and function of p8 [also known as Nupr1 and Com1 (Ree et al., 2000)], a small nuclear protein related to the HMGA1 family of chromatin architectural factors. We identified p8 in a gene-trap screen for loci induced in synchronized myoblasts (Sambasivan et al., 2008). However, this p300-binding phospho-protein has been implicated in the stress response (Vasseur et al., 2004), growth control (Malicet et al., 2003; Vasseur et al., 2002a; Vasseur et al., 1999), tumorigenesis (Vasseur et al., 2002b; Iovanna, 2002), metastasis (Ito et al., 2003) and cardiac hypertrophy (Goruppi et al., 2007). Here, we report a functional analysis of p8 in C2C12 myoblasts using RNA interference (RNAi), overexpression and yeast two-hybrid analysis. Our findings implicate p8 in the negative regulation of the cell cycle, and in the coordination of chromatin and transcriptional regulators to promote an early step in myogenic differentiation.

Suspension culture in methylcellulose medium generates homogeneous G0-arrested populations, whereas allowing the G0 cells to reattach leads to rapid and synchronous re-entry into G1 (Milasincic et al., 1996; Sachidanandan et al., 2002). This system facilitates the identification of G1-regulated genes, opening a window into the myogenic events that might be coupled to this phase. In a previous study, we used a gene-trap approach in combination with flow cytometry to screen for genes that were induced in G0 but suppressed during the subsequent S phase (Sambasivan et al., 2008; Sebastian et al., 2009). The nuclear protein p8 (Mallo et al., 1997) was one of 15 genes identified by this strategy. Expression of lacZ reporter RNA in the p8 gene-trap clone gtQ39 was induced >fourfold in G0-synchronized myoblasts as compared with asynchronously growing myoblasts, as was the endogenous p8 mRNA in parental C2C12 cells (Fig. 1A).

p8 encodes a small DNA-binding protein related to the HMGA1 family and has been implicated in the control of cell proliferation (Vasseur et al., 2002a; Malicet et al., 2003; Malicet et al., 2006). To address a possible role for p8 in growth control of muscle cells, we analyzed its expression during exit from reversible arrest (Fig. 1B). To monitor cell-cycle status, we used the S-phase-specific protein histone H2B – expression was absent in G0, induced at 14 hours of reactivation and peaked at 24 hours, consistent with timing of the S phase (Sachidanandan et al., 2002). By contrast, p8 mRNA, although expressed in G0, was strongly upregulated early during cell-cycle re-entry (2 hours), peaking at 6 hours (early G1) but returned to basal levels by 14 hours, well before the peak of S phase. This repression prior to S phase despite strong transient induction in G1 accounts for the recovery of p8 in the original screen. Notably, the peak of p8 expression precedes the induction of MyoD expression in G1 (Fig. 1C). Taken together, these results are consistent with a role for p8 during G0-G1.

To address the function of p8 in muscle cells, we used a knockdown strategy, using RNAi with short hairpin RNAs (shRNAs) (Yu et al., 2002) targeted against the p8 mRNA. Of the four shRNAs tested (p8sh1-p8sh4), growing myoblasts of the p8sh2 and p8sh4 pools exhibited a ∼60% reduction in the steady-state level of p8 mRNA compared with the vector control pool (Fig. 2A). The p8sh1 pool, which showed unperturbed levels of p8 mRNA, was used as an additional control.

Quantitative real-time reverse transcriptase (RT)-PCR analysis of a timecourse of activation from 2 to 24 hours in two independent pools showed suppression of the p8 transcript levels throughout the cell cycle but especially in mid-late G1 (Fig. 2B). p8 protein expression was also inhibited: Fig. 2C shows suppression in p8sh4-knockdown cells at the time of peak expression in late G1.

To monitor the consequences of p8 knockdown, the timing and extent of arrest in methylcellulose suspension culture and reactivation following reattachment were analyzed using flow cytometry. Growing cells of control (p8sh1) and knockdown [p8sh2 and p8sh4 pools and a clone derived from the p8sh4 pool (sh4 clone 12)] lines were compared with cells reactivated from arrest for different periods. Asynchronous populations of control and both p8-knockdown lines showed a similar cell-cycle profile (Fig. 2D). At 6 hours after reactivation from arrest, 90% of cells in both control and knockdown pools exhibited a G1 DNA content. However, at 16 hours of reactivation, whereas only 6% of control cells had entered S phase, significantly more knockdown cells had done so (14%, 25% and 50% in p8sh4-clone 12, p8sh4 pool and p8sh2 pool, respectively) (Fig. 2E). Control pools showed a substantial S-phase population only at 24 hours after reactivation. Taken together, these observations implicate p8 in the negative regulation of the cell cycle during G1.

To assess the effects of p8 knockdown on gene expression in G1, RNA was isolated from growing cells, G0-arrested cells and reactivated (mid-G1) cells of the control (p8sh1) and knockdown (p8sh4) pools. Northern blot analysis (Fig. 2F) showed a marked attenuation of p8-mRNA induction in the p8sh4 pool during G1. When compared with the respective G0 sample, p8 mRNA expression was induced tenfold in G1 in control cells, but not induced at all in the p8sh4 pool (quantified in Fig. 2G, lower panel). Thus, as in Fig. 2B, the effect of RNAi on p8 expression was strongest during the reactivation of myoblasts from G0 into G1.

Altering p8 expression affected both cell-cycle- and myogenic-marker expression. In control cells, S-phase-specific histone H2B expression was suppressed during G0 as expected and was yet-to-be induced at 6 hours of reactivation, consistent with the peak of S phase at >16 hours (Fig. 2F). By contrast, in knockdown cells (p8sh4 pool), histone mRNA was not as severely downregulated in G0 as in control cells and, by 6 hours of reactivation, knockdown cells had started to re-express histone mRNA, suggesting precocious entry into S-phase (quantified in Fig. 2G). Similar results were seen with the p8sh2 pool and p8sh4 clone 12 (not shown). Reduced p8 expression correlated with more rapid re-expression of histone mRNA, consistent with faster re-entry into S phase, implying a role for p8 in negative control of the G1-S transition in myoblasts. MyoD-mRNA induction in G1 was mildly reduced (twofold) in the p8-knockdown pool. Thus, a G1-induced transcript (MyoD) and an S-phase-specific transcript (histone) showed altered levels and timing in p8-knockdown myoblasts.

Myogenic differentiation is coupled to the G1 phase of the cell cycle (Clegg et al., 1987; Kitzmann et al., 1998). During differentiation, endogenous p8 expression was mildly induced at a stage when expression of myogenin was already activated (Fig. 3A), consistent with a role in myogenic-gene activation. Knockdown of p8 using either p8sh2 or p8sh4 led to reduced p8 transcript levels throughout the timecourse of differentiation (Fig. 3B). To explore the role of p8 in myogenesis, cells were incubated in low serum for 3 days. Whereas control cells expressed myosin heavy chain and fused to form multinucleated myotubes (fusion index 35-40%), p8-knockdown cells showed no fusion at all and only 9% of p8sh2 and 2.5% of p8sh4 cells expressed the muscle-specific myosin. Thus, compromising p8 expression also affects differentiation competence (Fig. 3C). To investigate the timing of the differentiation defect, control (shGFP) and knockdown (p8sh4) myoblasts were cultured in differentiation conditions for 12 hours and expression of early markers monitored. p8 protein levels were reduced in the p8sh4 cells (Fig. 3D). Interestingly, whereas MyoD protein levels were unaffected, induction of its target myogenin was strongly suppressed in knockdown cells. Expression of the CDKI p21, another target of MyoD with an important role in coupling arrest to differentiation, was also suppressed. However, expression of another CDKI, p27 (a marker of arrest that is not a direct MyoD target), was not affected, consistent with the ability of the p8-knockdown cells to exit the cell cycle (Fig. 2D).

Taken together, these observations are consistent with a role for p8 in regulating the ability of MyoD to activate its transcriptional targets.

To confirm the role of p8 in the control of myoblast growth and differentiation, we used ectopic expression of a human p8-FLAG construct in C2C12 myoblasts (Fig. 4). Whereas attenuation of p8 expression correlated with accelerated S-phase entry, overexpression of p8 completely inhibited BrdU incorporation, confirming that p8 negatively regulates the myoblast cell cycle. Whereas p21 was not affected, cells expressing p8-FLAG showed induction of p27, consistent with the inhibition of DNA synthesis. However, MyoD expression, which was seen in ∼50% of control EGFP-transfected myoblasts, was strongly repressed in p8-FLAG cells, and myogenin expression was not activated 24 hours after serum withdrawal. Thus, although endogenous levels of p8 are required for myogenesis (Fig. 3), sustained overexpression of this chromatin factor leads to arrest in a state that is refractory to differentiation.

p8 is related to the HMGA1 (formerly HMG I/Y) family (Encinar et al., 2001) of chromatin architectural proteins and binds to the HAT p300 in a complex that regulates glucagon gene expression (Hoffmeister et al., 2002). In addition to histones, p300 acetylates MyoD (Polesskaya et al., 2000) and serves as its co-activator (Yuan et al., 1996; Puri et al., 1997a), promoting MyoD function in differentiation. p300 also represses the transcription of Myc, an essential regulator of the G0-G1 transition (Kolli et al., 2001; Baluchamy et al., 2003). Thus, similar to p8, p300 promotes myogenesis and slows the cell cycle. To investigate whether the phenotype of p8-knockdown myoblasts (rapid S-phase entry and depressed MyoD function) is consistent with altered p300-dependent activities, we used three assays. First, we used quantitative RT-PCR to assess the levels of Myc expression during the synchronized cell cycle (Fig. 5A). In p8-knockdown cells, the expression of Myc mRNA was induced to higher levels and was sustained into late G1, consistent with the accelerated S-phase entry as well as inhibition of myogenesis (Miner and Wold, 1991).

Second, to determine the acetylation status of histones, we used western blotting with antibodies specific for histone modifications carried out by p300. A mild global reduction of histone acetylation [18% reduction of acetylated H3K18 (H3K18-Ac)] was observed in p8-knockdown cells, suggesting that p8 is a positive regulator of the HAT activity of p300 (Fig. 5B).

Finally, to determine the acetylation status of MyoD protein, we immunoprecipitated MyoD, followed by immunoblotting with a pan-acetyl lysine antibody (Fig. 5C). For equal amounts of MyoD protein, acetylation of MyoD in p8-knockdown myoblasts was only 38% of the level seen in control myoblasts. As p300 is known to acetylate MyoD as well as bind p8, our data are consistent with the model that lowering p8 expression compromises the function of p300 in post-translational modification of MyoD. Taken together, these data demonstrate that p8 regulates molecular events that are known targets of p300, consistent with p8 functioning through a p300-dependent mechanism.

To gain further insight into the function of p8, we used a yeast two-hybrid strategy to screen for proteins that interact with full-length mouse p8 (N.P., A.S. and J.D., unpublished). Among the candidates for a role in p8 function in myoblasts was p68 (Ddx5), another p300-binding protein (Fig. 6A). This DEAD-box RNA helicase is reported to complex not only with p300 (Rossow and Janknecht, 2003), but also with MyoD and SRA, a non-coding RNA (Caretti et al., 2006). Silencing of p68 suppresses myogenesis via mechanisms that are thought to involve altered chromatin remodeling by the ATP-dependent SWI-SNF Brg1 complex. Interestingly, p8 features in the list of genes that are downregulated by p68 RNAi (Caretti et al., 2006), suggesting that p68 is an upstream activator of p8. Indeed, p68 induction during G1 precedes the activation of p8 (data not shown).

To confirm the interaction between p8 and p68 we used two further tests. First, we used a mammalian two-hybrid assay. Full-length mouse p8 and p68 cDNAs were cloned into pBIND and pACT, respectively, and transfected into C2C12 cells along with the pluc5 reporter construct (Fig. 6B). Luciferase activity was induced twofold in the presence of both constructs, indicating a weak but reproducible interaction in mammalian cells. Second, we used an in vitro binding assay. Purified His-tagged mouse p8 protein was able to pull down p68 from C2C12 myoblast lysate (Fig. 6C). We also confirmed that p8 interacts with p300 in this assay (Fig. 6C). Notably, neither p68 nor p300 expression was affected in the p8-knockdown cells (Fig. 6D).

The ability of p8 to affect MyoD acetylation as well as to interact with two important chromatin regulators that are also known to bind MyoD suggested the possibility of a direct interaction between p8 and MyoD. Therefore, we tested whether p8 could directly bind MyoD. As with p68 and p300, His-tagged p8 could specifically pull down MyoD from myoblast lysates (Fig. 6C). The specificity of the interaction between p8 and either MyoD, p68 or p300 was underscored by the inability of His-p8 to pull down histone H3 (Fig. 6C).

To confirm the interaction between p8 and MyoD, we coexpressed MyoD-YFP and FLAG-p8 in HEK293 cells. Immunoprecipitation of p8 with anti-FLAG also recovered MyoD, providing further evidence for their interaction. Thus, p8 interacts with three proteins already reported to individually interact with each other – p300, p68 and MyoD.

Direct interaction of p8 with three chromatin-binding proteins and transcriptional activators, two of which (MyoD and p300) are known to bind the myogenin promoter, suggested that p8 might also be associated at this element. The myogenin promoter is a key genomic target of mechanisms that regulate differentiation. Interestingly, not only were MyoD and p300 detected at this site, but p8 itself and the RNA helicase p68 also bound to this early differentiation promoter (Fig. 7). Furthermore, the association of all four proteins was reduced in p8-knockdown cells. Whereas reduced association of p8 could be attributed to lower levels of p8 protein in knockdown cells (Fig. 2C; Fig. 3D), reduced association of the three binding partners of p8 is not due to lower expression levels (Fig. 3D; Fig. 6D). Thus, occupancy of the myogenin promoter by all three p8-interacting proteins – MyoD, p300 and p68 – was dependent on p8 expression and chromatin association.

Taken together, our findings demonstrate that, in myoblasts, p8 physically binds to three other transcriptional modulators and co-activators with a common chromatin target – the early muscle-specific myogenin promoter – and regulates not only their chromatin association but also the post-translational modification and function of MyoD, a key activator of this promoter. Thus, p8 is a cell-cycle-regulated molecule that orchestrates the functioning of an important early regulatory hub during myogenic differentiation (Fig. 8).

In this study, we investigated the function of p8, a small chromatin-binding protein whose expression peaks in G1 (Sambasivan et al., 2008). We establish that p8 inhibits cell-cycle progression and regulates myogenic differentiation. We confirm that p8 binds the HAT p300, and report two new interactions with proteins also known to bind p300: the RNA helicase p68 and the myogenic regulator MyoD. We show that p8 is required for MyoD acetylation, itself associates with the myogenin promoter, and regulates association of MyoD, p300 and p68 with this promoter. Our findings suggest that p8 orchestrates the action of chromatin factors that converge to regulate an early muscle-specific promoter, and represents a new node in which control of myogenic differentiation intersects with the cell cycle.

p8 has been previously implicated in growth control: knockout fibroblasts show reduced cell-cycle time (Vasseur et al., 2002a; Malicet et al., 2003), suggesting that this chromatin-binding protein inhibits cell proliferation. In myoblasts, as in fibroblasts, p8 acts as a brake on the cell cycle. Altering p8 expression in opposite ways has reciprocal effects on DNA synthesis: reducing p8 expression accelerated the kinetics of G1 progression to S phase, whereas ectopic expression of p8 led to arrest in G0-G1. These results suggest that not only induction of p8 in early-mid G1, but its repression prior to S phase, are essential for normal cell-cycle kinetics. The effects of p8 on G1 kinetics might be mediated by induction of the G1 CDKIs p27 and p21, and by negative control of Myc, a major regulator of progression that acts on a large number of cell-cycle promoters (Gartel and Shchors, 2003; Wanzel et al., 2003). Sustained expression of Myc is also known to suppress differentiation (Miner and Wold, 1991).

The `shrinking' of G1 caused by reduced p8 expression is associated with a loss of tissue-specific gene activation, suggesting that p8 might control events in G1 that bridge the programs of proliferation and differentiation. MyoD plays a key role at this intersection by two mechanisms: transcriptional induction of cell-cycle inhibitors (Halevy et al., 1995; Magenta et al., 2003) and myogenic activators, as well as direct binding of Cdk4 to inhibit phosphorylation of Rb (Zhang et al., 1999b), promoting differentiation-coupled arrest. In this context, our finding that p8 not only negatively regulates the G1-S transition but also positively modulates MyoD function suggests participation in the mechanism that restricts myogenic differentiation to the G1 phase.

p8-knockdown myoblasts failed to differentiate owing to the absence of key early differentiation regulators, namely myogenin and p21. However, p8 overexpression in growing myoblasts did not enhance the myogenic program. This is probably an indirect consequence of arrest in early G1 prior to MyoD activation, blocking the development of differentiation competence.

p8 shares 35% amino acid identity with the HMGA1 (formerly HMGI/Y) chromatin architectural factors (Encinar et al., 2001) and can regulate the activity of different transcription factors, such as Smad (Garcia-Montero et al., 2001), p53 (Clark et al., 2008), Jun and other established AP1 effectors (Goruppi et al., 2007). p8 also binds to p300 and Pax2-transactivation domain interacting protein (PTIP) to regulate the activity of Pax2A and Pax2B on the glucagon promoter (Hoffmeister et al., 2002).

In myoblasts, interactions with other chromatin proteins also point to a mechanism by which p8 influences tissue-specific gene expression. We show that p8 binds three proteins: the HAT p300, the RNA helicase p68 and MyoD. The role of p300 in the control of myogenesis and the cell cycle is well established. p300 influences two activities that are de-regulated in p8-knockdown myoblasts: MyoD function (both by acetylating MyoD and by acting as its co-activator via HAT activity) (Polesskaya et al., 2001; Yuan et al., 1996; Puri et al., 1997a; Sartorelli et al., 1997; Magenta et al., 2003), and Myc expression (Kolli et al., 2001; Baluchamy et al., 2003). Our study confirms the direct interaction of p8 with p300 and establishes a crucial role for p8 in p300 function, because common as well as cell-type-specific p300-dependent activities are altered when p8 expression is compromised.

The identification of the RNA helicase p68 as a partner of p8 in an unbiased interaction screen expands the scope of this small chromatin-binding factor, because interactions are reported between p68 and p300 (Rossow and Janknecht, 2003), as well as between p68 and MyoD (Caretti et al., 2006). In conjunction with the non-coding RNA SRA, p68 is essential for muscle differentiation. Although ubiquitously expressed, both p300 and p68 are particularly relevant to myogenesis because they enhance MyoD function. Pharmacological inhibition or genetic ablation of p300 HAT activity (Polesskaya et al., 2001; Roth et al., 2003) or silencing of p68 (Caretti et al., 2006) all phenocopy the p8 knockdown: the expression of myogenin and myosin heavy chain as well as myoblast fusion competence are compromised, much as when MyoD itself is inactivated (Yablonka-Reuveni, et al., 1999).

The mechanism by which p8 affects differentiation probably involves reduced acetylation by p300 of MyoD; p300 modifies three lysines near the bHLH domain of this protein. Acetylation leads to increased binding to p300 and enhances MyoD activity. Similar to MyoD, p8 is acetylated by p300, which leads to enhanced transcriptional activity of associated tissue-specific factors such as Pax2 (Hoffmeister et al., 2002). Analysis of p8 secondary structure predicts a HLH motif (Goruppi et al., 2007) towards the C-terminus (positions 46-71), suggesting the potential to dimerize with other HLH proteins, including MyoD (Murre et al., 1989; Benezra et al., 1990). However, the importance of the HLH motif is unresolved because our preliminary studies suggest that overexpression of a mutant p8 lacking helix 2 mimics wild-type p8 in inhibiting the cell cycle as well as MyoD and myogenin expression (S.C., unpublished).

p8 might not interact with DNA directly but probably associates with chromatin via transcriptional activators and co-activators (Goruppi et al., 2007). In addition, p8 knockdown might compromise the function of p300 and/or p68 on other targets. For example, p300 acetylates Rb (Nguyen et al., 2004) and MEF2 (Ma et al., 2005), both of which are crucial for myogenic differentiation. Altered chromatin remodeling might also play a role, because MyoD normally recruits p300 to its target promoters (Giacinti et al., 2006; Wilson and Rotwein, 2006) and p68 interacts with HDAC1 (Wilson et al., 2004), influencing the recruitment of the Swi-Snf chromatin-remodeling protein Brg1 and the transcription machinery onto MyoD target promoters (Caretti et al., 2006).

Our findings suggest that four proteins – p8, p68, p300 and MyoD – participate in a common mechanism to modulate myogenesis via chromatin remodeling at the myogenin promoter. As shown in this report for p8, depleting either p300, p68 or MyoD (Baluchamy et al., 2003; Kolli et al., 2001; Caretti et al., 2006; Rudnicki et al., 1993; Yablonka-Reuveni et al., 1999) also leads to rapid proliferation and defective differentiation. Thus, p300, p68, MyoD and p8 are all anti-proliferative and pro-myogenic, interact with each other and with a common chromatin target, and similar effects on the cell cycle and muscle differentiation seem to result by compromising each of their functions individually. Taken together, these findings suggest that the chromatin architectural protein p8 regulates myogenic differentiation via the effects of the chromatin modulators p300 HAT and p68, and the myogenic determination factor MyoD. Because p300, p68 and p8 are all expressed in G0 myoblasts (Sindhu Subramaniam and J.D., unpublished) but MyoD is absent (Milasincic et al., 1996; Kitzmann et al., 1998; Sachidanandan et al., 2002), the composition of p8-containing complexes might vary in quiescent, proliferating and differentiating muscle cells.

p8-overexpressing cells could be staged as G0 or early G1, because MyoD expression – characteristic of mid-G1 – is absent. Therefore, arrest of the cell cycle by p8 led to an uncoupling of myogenesis from cell-cycle exit. p8 seems to regulate the pace of G1 progression such that MyoD expression is activated appropriately and MyoD protein is sufficiently acetylated.

According to the `G1 model' of cell-cycle regulation, decisions to divide, arrest, differentiate or die are taken during this phase (Pardee, 1989; Clegg et al., 1987; Riddle et al., 1979), when environmental and intrinsic cues are assessed and integrated towards an appropriate response. Cells lose their responsiveness to external signals during the Rb-controlled restriction point at the G1-S boundary (Planas-Silva and Weinberg, 1997). Classical studies indicated that myogenic differentiation is only initiated during the G1 phase (Clegg et al., 1987), whereas recent studies provide evidence for a control point in G1 in which cell-cycle progression and myogenic differentiation are linked (Kitzmann and Fernandez, 2001; Zhang et al., 1999a). However, although MyoD itself has been implicated in this link, the molecular mechanisms by which this restriction occurs are not clear. Our findings suggest that p8 contributes to the coupling of differentiation to the G1 phase of the cell cycle via effects on post-translational modifications that affect the transcriptional function of MyoD and the assembly of a cohort of anti-proliferative and pro-myogenic regulators on an early muscle-specific promoter. The identification of p8 as an interacting partner of MyoD, and its association with p68 as well as p300, expands the list of molecular players at this regulatory node, providing a framework in which to dissect its mechanism.

C2C12 myoblasts (Yaffe and Saxel, 1977; Blau et al., 1983) were obtained from Helen Blau (Stanford, CA) and a sub-clone A2 (Sachidanandan et al., 2002) used in all experiments. Myoblasts were maintained in growth medium (GM: DMEM with 20% FBS). Differentiation was induced by incubating cultures in differentiation medium (DM: DMEM with 2% horse serum), replaced daily for 3-5 days. Differentiation is expressed as the fusion index (calculated as % of total nuclei present in myotubes of >two nuclei).

Suspension culture was performed as described (Milasincic et al., 1996; Sachidanandan et al., 2002). Briefly, sub-confluent cultures were harvested and cultured as a single cell suspension (10⁵ cells/ml) in DMEM containing 1.3% methyl cellulose and 20% FBS. After 48 hours (>98% of cells arrested in G0), cells were harvested by dilution and centrifugation. G0 cells were reactivated into the cell cycle by replating in GM and harvested 6-24 hours later. Upon reattachment, G0 myoblasts synchronously enter G1 at ∼4-6 hours, with a peak of S phase at 24 hours (Sachidanandan et al., 2002).

Cells plated on coverslips were fixed with 3.5% formaldehyde/PBS and permeabilized in PBS/0.2% Tween-20. Primary antibodies were diluted in PBS/10% HS/0.02% Tween: anti-FLAG (Sigma) 1:1000; anti-MyoD (DAKO), anti-p21 (Santa Cruz), anti-p27 (Santa Cruz), anti-BrdU (BD Biosciences), anti-myogenin (Santa Cruz) all at 1:100; anti-myosin (A4-1025) 1:10 (Blau et al., 1983). Detection of incorporated BrdU was performed after denaturation of DNA using 0.4 N HCl, 0.5% Tween, 0.5% Triton X-100 as described (Sachidanandan et al., 2002). Secondary antibody was goat anti-mouse Alexa Fluor 488 or 594 (Molecular Probes), 1:500. Secondary-antibody controls were negative; no cross-reactivity of secondary reagents was detected. Nuclei were detected with Hoechst 33342 (1 μg/ml). Staining was recorded on a CCD camera using an Olympus microscope equipped with epifluorescence or on a Zeiss LSM510 Meta confocal microscope. Images were assembled using Adobe Photoshop 6.0.

Northern blot analysis was performed as described (Sachidanandan et al., 2002) using a probe spanning nucleotides 147-550 of mouse p8 mRNA. L-Process and Image Gauge programs (Fuji) were used to quantify background-subtracted phosphor-imager signals.

Western blot analysis was performed as described (Sachidanandan et al., 2002). Antibodies were diluted in blocking buffer: anti-MyoD (Santa Cruz) 1:400; anti-myogenin (Santa Cruz) 1:500; anti-p27 (BD Bioscience) 1:3000; anti-p21 1:1000; p8 polyclonal antisera (kind gift of Juan Iovanna, INSERM, Marseille, France) 1:200. HRP-conjugated secondary antibody (Bangalore Genei 1:10,000) was detected using ECL (Amersham).

Potential targets for shRNA-mediated silencing of p8 expression were identified using OligoRetriever (http://www.cshl.org/public/SCIENCE/hannon.html). Four different shRNA hairpins were designed, each containing 21 b antisense and sense sequences separated by a loop (5′-GAAGC-3′). For each shRNA construct, complementary oligos (Microsynth, Switzerland) were annealed and ligated into the mU6 promoter vector (Yu et al., 2002).

C2C12 myoblasts were transfected (Lipofectamine, Invitrogen) with each of the shRNA constructs plus a zeocin selection marker pKA23 at ratio of 4:1, and selected in 200 μg/ml zeocin to generate stable pools. Individual clones were isolated by ring cloning.

p8-knockdown shRNAs: P8SH-2a: 5′-TTTGGGCTGTCTTCCTAGCTCTGCCGAGCGGCAGAGCATGGAAGACAGCCTTTTT-3′; P8SH-2b: 5′-CTAGAAAAAGGCTGTCTTCCATGCTCTGCCGCTTCGGCAGAGCTAGGAAGACAGCC-3′; P8SH-4a: 5′-TTTGGGGCCAGGCTGTACTGATCATGAAGCATGATCTGTACACCCTGGCCCTTTTT-3′; P8SH-4b: 5′-CTAGAAAAAGGGCCAGGGTGTACAGATCATGCTTCATGATCAGTACAGCCTGGCCC-3′.

Control shRNAs: P8SH1a: 5′-TTTGTCTTGCCTGTGTCTCCTTGTCGAAGCGACAAGGACTCACAGGCAAGATTTTT-3′; P8SH1b: 5′-CTAGAAAAATCTTGCCTGGAGTCCTTGTCGCTTCGACAAGGAGACACAGGCAAGA-3′; GFP - SH1a: 5′-TTTGAACTTCAAGGTCCGCCACAACGAAGCGTTTTGGCGGACCTTGAAATTTTTTT-3′; GFPSH1b: 5′-CTAGAAAAAAATTTCAAGGTCCGCCAAAACGCTTCGTTGTGGCGGACCTTGAAGTT-3′.

Flow cytometry was performed on a FACSCaliber (BD Biosciences), using CelQuest software as described (Sachidanandan et al., 2002).

Control or RNAi myoblasts grown to high density were incubated overnight in DM containing sodium butyrate (1 mM) to inhibit histone decactylase activity. Nuclei were isolated from ∼10⁷ cells using NP40 lysis buffer (10 mM Tris, pH 7.5, 5 mM MgCl₂, 10 mM NaCl, 1% NP40) containing protease inhibitors and extracted by gentle Dounce homogenization in Dignam C buffer (20 mM HEPES, pH 8, 10% glycerol, 420 mM NaCl, 1.5 mM MgCl₂, 0.1 mM EDTA). Cleared nuclear protein extracts (equal protein) were diluted in modified RIPA buffer, incubated with 5 μg anti-MyoD monoclonal antibody (Dako) overnight, immune complexes captured using protein-G beads, washed with cold PBS + 0.5% Triton X-100 and pelleted.

IP pellets were solubilized in 2× Laemmli sample buffer and run on 10% SDS-PAGE. Antibodies were diluted in blocking buffer: MyoD polyclonal (Santa Cruz) 1:400, pan-acetyl lysine monoclonal (Upstate) 1:2000.

Full-length human p8 cDNA was obtained from A549 cells using RT-PCR using primers with BamH1 and Xho1 restriction sites (5′-AATGACGGATCCATGGCCACCTTCCCACCAGCA-3′ and 5′-GAACTACTCGAGTCAGCGCCGTGCCCCTCGCT-3′) and cloned into pCMV-2b to generate an N-terminal FLAG-tag.

Total RNA (1 μg) isolated from stable pools (p8sh or GFP-sh) was used to generate cDNA (Clontech). 2 μl of cDNA (diluted 1:5) were mixed with 10 μl of SYBR Green PCR Master Mix (Applied Biosystems) and analyzed in triplicate using a 7900HT cycler (Applied Biosystems). Normalized fold differences of cycle thresholds [2^{–(–ΔΔCt)}] of Myc and p8 amplicons were calculated relative to a control L7 amplicon; dissociation curves and sequencing were used to verify amplicons. Primers were: Myc: 5′-GCGCAAAGACAGCACCA-3′ and 5′-GCGAGCTGCTGTCGTTGA-3′; L7: 5′-GGAGCTCATCTATGAGAAGGC-3′ and 5′-AAGACGAAGGAGCTGCAGAAC-3′; p8: 5′-AGGACCTAGGCCTGCTTGAT-3′ and 5′-CTCTGCTTCTTGCTCCCATC-3′.

Mouse p8 cDNA (coding) was amplified from mRNA of G0 myoblasts by RT-PCR using primers with EcoRI and SalI restriction sites (5′-ATCGAATTCGGCATAATGGCCACC-3′ and reverse 5′-ATGTCGACGTGCTGTCACTGCTGT-3′), and cloned into the pGBKT7 vector as bait for screening a mouse 7-day embryo Matchmaker cDNA library in pACT2 (Clontech), in Saccharomyces cerevisiae PJ694A, according to the manufacturer's instructions. Transformants passing three rounds of selection on dropout plates (Trp^–Leu^–Ade^–) were re-tested on Trp^–Leu^–Xgal⁺ plates to confirm activation of the β-galactosidase reporter. Plasmids were isolated from positive yeast clones, propagated in E. coli DH5α, verified by re-transformation into yeast with the original p8-bait plasmid or control plasmids, selected as above, and sequenced.

Mouse p8 cDNA cloned into pBIND (Clontech; using the primers 5′-GCTGCACGGATCCTAATGGCCACCTTGCCACCA-3′ and 5′-GCATATTCTAGATGCTTGCACTGCTGTACGATT-3′) and full-length mouse p68 cDNA cloned into pACT using the primers 5′-GCTGCACGGATCCGCATGTCGAGTTATTCTAGTGAC-3′ and 5′-GCATATTCTAGATTGAGAATACCCTGTTGGCATG-3′) were introduced into C2C12 cells along with p5luc reporter and luciferase assays performed on lysates prepared 48 hours after transfection.

Mouse p8 cDNA was cloned into the EcoRI-SalI sites of pET28 (Novagen) and the fusion protein with an N-terminal His-tag (His⁶mp8) purified from E. coli using Ni²⁺-NTA resin (Qiagen). 50 μl bed volume of nickel-agarose bound to His⁶mp8 was incubated in 100 μl of pulldown buffer (20 mM HEPES/KOH, pH 7.5, 100 mM KCl, 5 mM MgCl2, 0.5 mM EDTA, 0.05% NP-40, 1 mM DTT, 0.02% BSA, and protease inhibitors), with 200 μg of C2C12 myoblast extract prepared by salt extraction of isolated nuclei using Dignam C buffer as detailed above. After 3 hours of incubation at 4°C, the nickel-agarose beads were washed 5× with PBS before elution of bound proteins with Laemmli sample buffer. One tenth of eluted material was subjected to western blotting with antibodies against p300 (clone RW128; Upstate), p68 (pAb204; Upstate), MyoD (polyclonal, Santa Cruz); His-p8 (anti-His tag, Santa Cruz) or the control Histone H3 (polyclonal, Abcam).

p8-FLAG and MyoD-YFP were co-transfected into HEK293 cells and cell extracts subjected to immunoprecipitation with anti-FLAG antibody and western blotting with anti-MyoD.

Control and p8 RNAi myoblasts were incubated in DM for 24 hours, chromatin isolated, cross linked and subject to immunoprecipitation with antibodies against p8, p68, p300, MyoD or H3. Antibodies to p300, p68, MyoD and H3 are as above; for p8, a new polyclonal antibody raised against mouse p8 was used. The myogenin promoter (–200 fragment containing the E box) was PCR amplified [primers: 5′-GAATCACATGTAATCCACTGGA-3′ and 5′-ACGCCAACTGCTGGGTGCCA-3′, and 5′-AGAGGGAAGGGGAATCACAT-3′ and 5′-CATTTAAACCCTCCCTGCTG-3′] from DNA recovered after reversal of cross-linking, using SYBR-green QPCR reactions in an ABI 7900HT real-time cycler.